CORAS

Data Scopes

Control exactly which data Gary can access in chat — configure once, apply across your team.

A data scope is an AI configuration that defines which fields, in which lists, the AI chat is allowed to access. Linking a data scope to a Smart Agent means users automatically get the right data context when they open a chat with that agent — no manual setup required.

Note: In "CORAS Light", data scopes operate in search mode. The Similarity and Predictions modes available in full CORAS are not exposed in "CORAS Light."

Create a Data Scope

  1. Open Admin Area and go to Data Scopes.

  2. Click Add Data Scope.

  3. Enter a name and select the lists and fields you want Gary to access.

  4. Click Save.

After creating a data scope, link it to a Smart Agent so the right scope is applied automatically whenever a user opens that agent in chat.

Manage Data Scopes

From the Data Scopes list, you can take the following actions on any scope:

  • Edit — open a scope and update its lists, fields, or name.

  • Sync — refreshes the data scope from its source. Use this when underlying data has changed and you want Gary's answers to reflect the latest state.

  • Build — processes and indexes the data scope. Run this after creating a scope or making significant changes before users start querying.

  • Delete — removes the scope permanently. Agents linked to a deleted scope lose their data access until you link a new one.

Data Scopes in CORAS Light

In CORAS Light, Gary operates in search-only mode — Gary can retrieve and surface information from your data sources, but cannot write output back to records.

What this means for Data Scope configuration:

Data Scopes in CORAS Light control what Gary can search and return in responses. Gary reads from the sources and fields included in the scope. Write-to-field actions are not available in CORAS Light regardless of scope settings.

Configuring a Data Scope for a CORAS Light tenant:

The setup steps are the same as a standard tenant — define the scope, select sources, and link it to an agent or AI Config. The difference is in what Gary can do with that scope: search and surface information only.

If your team needs Gary to write to records:

Writing to records requires the full CORAS platform. Contact your CORAS administrator or support contact to discuss your tenant configuration.

When to Sync vs. When to Build

Sync and Build do different jobs.

Sync updates the scope's understanding of your list structure — new fields, renamed fields, removed fields. Run it after you make changes to the lists included in the scope.

Build processes and indexes the actual data so Gary can query it accurately. Run it after creating a scope for the first time, and after significant data changes like a bulk import or major record updates.

In most cases: Sync first when structure changes, Build after when data changes.

For End Users: Starting a Data Chat

When you open a new Data Chat in CORAS Light, you'll see the Data Scopes your admin has made available to you. Select the one that matches the data you need, and Gary will draw from that dataset throughout your conversation.

If the scope you need isn't in the list, contact your administrator — Data Scopes are configured and assigned by admins.

Get Help

Contact your Customer Success Manager or visit the Customer Resource Center for additional support with Data Scopes and AI Settings.

Previous

AI Skills